Vibrio fischeri Possesses Xds and Dns Nucleases That Differentially Influence Phosphate Scavenging, Aggregation, Competence, and Symbiotic Colonization of Squid

Abstract

Cells of Vibrio fischeri colonize the light organ of Euprymna scolopes, providing the squid bioluminescence in exchange for nutrients and protection. The bacteria encounter DNA-rich mucus throughout their transition to a symbiotic lifestyle, leading us to hypothesize a role for nuclease activity in the colonization process. In support of this, we detected abundant extracellular nuclease activity in growing cells of V. fischeri. To discover the gene(s) responsible for this activity, we screened a V. fischeri transposon mutant library for nuclease-deficient strains. Interestingly, only one strain, whose transposon insertion mapped to nuclease gene VF_1451, showed complete loss of nuclease activity in our screens. A database search revealed that VF_1451 is homologous to the nuclease-encoding gene xds in Vibrio cholerae. However, V. fischeri strains lacking xds eventually revealed slight nuclease activity on plates after 72 h. This led us to hypothesize that a second secreted nuclease, identified through a database search as VF_0437, a homolog of V. cholerae dns, might be responsible for the residual nuclease activity. Here, we show that Xds and/or Dns are involved in essential aspects of V. fischeri biology, including natural transformation, aggregation, and phosphate scavenging. Furthermore, strains lacking either nuclease were outcompeted by the wild type for squid colonization. Understanding the specific role of nuclease activity in the squid colonization process represents an intriguing area of future research. IMPORTANCE From soil and water to host-associated secretions such as mucus, environments that bacteria inhabit are awash in DNA. Extracellular DNA (eDNA) is a nutritious resource that microbes dedicate significant energy to exploit. Calcium binds eDNA to promote cell-cell aggregation and horizontal gene transfer. eDNA hydrolysis impacts construction of and dispersal from biofilms. Strategies in which pathogens use nucleases to avoid phagocytosis or disseminate by degrading host secretions are well documented; significantly less is known about nucleases in mutualistic associations. This study describes the role of nucleases in the mutualism between V. fischeri and its squid host, Euprymna scolopes. We find that nuclease activity is an important determinant of colonization in V. fischeri, broadening our understanding of how microbes establish and maintain beneficial associations.

Keywords: Euprymna scolopes; Vibrio fischeri; biofilm; genetic competence; nuclease; symbiosis.

FIG 1

V. fischeri possesses two putative secreted nucleases that are homologs of Xds and Dns nucleases in V. cholerae. (A) The Xds proteins in V. fischeri ES114 and V. cholerae VC0395 share two functional domains, lamin tail domain (LTD; diagonal stripes) and the endonuclease/exonuclease/phosphatase family (stippled). (B) The Dns proteins share an endonuclease I functional domain (horizontal stripes). The numbers listed above the protein cartoon represent amino acid number. (C and D) Genes flanking the xds homologs are different (C), while the genes flanking dns in both species are homologs (D).

FIG 2

Nuclease activity in strains of V. fischeri. (A) Strains lacking xds (i.e., xds and xds-dns) were unable to degrade DNA in DNase agar after 24 h at 28°C, but expression of xds in trans in the xds single mutant (xds⁺) restored nuclease activity. (B) After 72 h, a halo around the xds mutant, but not the xds dns double mutant, suggested that Dns might eventually compensate for lack of Xds. (C) Nuclease activity in cells grown with shaking in LBS to an OD₆₀₀ of either 0.5 or 1.0. Cell-free culture supernatants were mixed with DNaseAlert reagents to measure nuclease activity. Each bar represents the average of 5 replicates (with standard error); these results are representative of three separate experiments. Different letters indicate statistically significant differences based on analysis of variance (ANOVA) (P < 0.05).

FIG 3

Effect of EDTA and phosphate on V. fischeri DNase activity. (A) Wild-type (ES114) cells were grown with shaking in LBS to an OD₆₀₀ of 0.5 and 1.0. Cell-free supernatants with and without added EDTA (10 mmol/L) were mixed with DNaseAlert reagent to measure nuclease activity. Each bar represents the average of 5 replicates (with standard error); these results are representative of three separate experiments. (B) Cells were grown in minimal medium containing either 0.3 mM or 3 mM phosphate and then assayed for nuclease activity as described previously. Each bar represents the average of 5 replicates from three experiments (with standard error). Different letters indicate significant difference based on ANOVA. (P < 0.05).

FIG 4

Degradation of eDNA by V. fischeri strains. Strains were cultured in TMM to an OD₆₀₀ of approximately 2.0. Either sterile TMM or cell-free supernatant from each strain was mixed with ~0.2 μg of PCR DNA (linear) (A) or 1 μg of plasmid (pUC19) (B) incubated at 28°C for the times (in minutes) indicated and then visualized on a 1% agarose gel.

FIG 5

Secreted nucleases are essential for optimal growth when eDNA is the only source of phosphate. Cells were grown at 28°C with shaking in TMM containing sheared salmon sperm DNA (1.0 mg mL⁻¹) (A) or K₂HPO₄ (B) as the sole phosphate source or no phosphate (C). Error bars represent averages (± standard error) for three separate trials.

FIG 6

Xds is required to digest DNA in mucin. The absence of a halo around the xds mutants (i.e., xds, xds-dns) suggests that, for at least 24 h, no other nuclease participates in the breakdown of DNA in mucin. Complementation of xds in trans (i.e., xds⁺) partially restored the ability to digest mucin. Two versions of double mutant (i.e., xds-dns and xds-dns/KV8865) showed no difference in nuclease activity on mucin plates.

FIG 7

Dns mutant cells display a hyperaggregation phenotype. Cells were grown in minimal medium without shaking for up to 1 week and then gently mixed to visualize aggregation. Only strains lacking dns (i.e., dns and xds-dns) showed an aggregation phenotype.

FIG 8

The dns mutant is hypertransformable. Strains ES114, dns (KV8807), xds (KV8811), xds-dns (KV8865), and dns⁺ (KV10274) carrying a tfoX expression plasmid with either kanamycin (A) or chloramphenicol resistance (B) were grown at 28°C with shaking in TMM to an OD₆₀₀ of approximately 0.5. Aliquots of each culture were exposed to purified DNA containing an antibiotic resistance marker. Cells were plated on selective media, transformants were counted to estimate transformation frequency, and an ANOVA test was performed assuming normal distribution. Dotted lines indicate limit of detection. *, P < 0.05; **, P < 0.005; ****, P < 0.00005; ns, not significant.

FIG 9

Single-strain squid colonization. (A) Closed squares, closed triangles, and open circles represent bioluminescence emission per squid infected with either xds, dns, or the wild-type strain ES114, respectively. Red lines represent average relative light units (RLU) per squid over the 48-h time period. The number of colonized squid was determined by comparing RLU emission between aposymbiotic squid and the test strains. (B) Nuclease is not required for normal growth or luminescence in culture. All results shown are typical for three independent trials.

FIG 10

Nucleases are required for optimal squid colonization. Juvenile squid were exposed to an inoculum containing a 1:1 ratio of wild-type V. fischeri and either the xds (A) or dns (B) mutant. At 48 h postinfection, the ratio of mutant to wild-type bacteria in each light organ was assessed by plating squid light organs. The results are expressed as the relative competitive index (RCI). An RCI above 1.0 indicates that the mutant was dominant. These results are representative of three separate trials.